Modelled gridded population estimates for the Kasaï-Oriental Province in the Democratic Republic of Congo (2024) version 4.2

Content

This repository contains the input data and scripts used to create the modeled gridded population estimates for Kasaï-Oriental Province in the Democratic Republic of Congo. It also includes the grid-cell posterior distributions and scripts to aggregate them within user-defined geographic boundaries.

 In particular, this repository contains two compressed files (.zip):

1. population_estimates.zip

Includes raster files (.tif) with summaries of population count posterior predictions at the grid-cell level, specifically the mean, median, lower credible interval, and upper credible interval.

Includes spatial files (.gpkg) with summaries of population count posterior predictions at the health-area and health-zone levels, specifically the mean, median, lower credible interval, and upper credible interval.

2. population_model.zip

This directory comprises five subdirectories with scripts, input data, and output data necessary to replicate the population model:

01_model_stan: Contains the Stan model, input data, and an R script (01_model_stan.R) with a function to run the model.

02_model_run: Includes an R script (02_model_run.R) for running the model, along with output data.

03_model_evaluate: Features a Quarto report template (03_model_evaluate.qmd) and model evaluation summary files(.pdf).

04_predict_posterior: Provides R scripts (04_predict_posterior.R and 04_predict_run.R) for generating predictions, along with input and output data, namely the posterior predictions files (.rds).

05_aggregate_posterior: Contains R scripts (05_aggregate_posterior.R and 05_aggregate_run.R) and associated input and output data, namely the population count posterior summaries as presented in the file population_estimates.zip .

The work was carried out in R (version 4.4.0), with the packages tidyverse (version 2.0.0), terra (version 1.7-78), sf (version 1.0-16), furrr (version 0.3.1), doParallel (version 1.0.17), foreach (version 1.5.2), rstudioapi (version 0.16.0), and rstan (version 2.32.6), on macOS Sequoia (version 15.1.1). While the scripts are designed to be portable, minor adjustments may be required for compatibility with other operating systems.

Important

This version includes changes in the STAN model 10h_survey_survey_covariate_building_random_effect_hierarchy_building_covariate_density_fixed_effect_hierarchy_density.stan. Consequentely, all the files generated in the previous versions are now changed.

For inquiries regarding the model and the data, please contact Gianluca Boo at [e-mail].